From: Russ Allbery (rra@stanford.edu)
Date: Fri Jul 02 2004 - 14:42:58 CDT
Charles Lindsey <chl@clerew.man.ac.uk> writes:
> Anyway, for anything bad to happen, two things are required:
I think you're confused about the problem that we're worrying about. All
that's required for something bad to happen is for a user to write a
Subject that begins with "cmsg" without intending it to be a control
message. The article will then mysteriously disappear on servers that
consider that to be a control message.
In order to avoid this transitional breakage, we need to make it clear
that articles with a subject beginning with "cmsg" that are not control
messages with a matching Control header will encounter real-life
interoperability problems and therefore are a bad idea until support for
that method of specifying a control message goes the way of .ctl
pseudo-groups.
> Just say that Subject: cmsg MUST NOT be interpreted as a control message
> (which is what the current draft says) and leave it at that.
I think going from MUST to MUST NOT without any transition whatsoever is a
very iffy thing to do, and in practice won't result in a clean transition.
-- Russ Allbery (rra@stanford.edu) <http://www.eyrie.org/~eagle/>